Looking at the patch I didn't like that duplicated the page scan in a second function without refactoring out the first scan. I think this comes from the usual noncommiter-itus of trying to modify the existing code as little as possible. At least that's the problem I've had which led to things like this. Instead try to figure how you would have written it if you were writing it from scratch.
If it's convenient to have a function to handle the scan then use it for both scans. You could have a flag that indicates it should abort after the first freeze. I think it would be simpler to have a return value indicating the oldest transaction found and then just call it again if that's old enough for the opportunistic threshold. -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers
